Potential Crosstalk between Sonic Hedgehog-Wingless-related Integration Site Signaling and Neurovascular Molecules: Implications for Blood Brain Barrier Integrity in Autism Spectrum Disorder

Abstract

Autism Spectrum Disorder (ASD) is a neurodevelopmental disease originating from combined genetic and environmental factors. Post-mortem human studies and some animal ASD models have shown brain neuroinflammation, oxidative stress and changes in blood brain barrier (BBB) integrity. However, the signaling pathways leading to these inflammatory findings and vascular alterations are currently unclear. The BBB plays a critical role in controlling brain homeostasis and immune response. Its dysfunction can result from developmental genetic abnormalities or from neuroinflammatory processes. In this review we explore the role of the Sonic Hedgehog/Wingless-related integration site (Shh/Wnt) pathways in neurodevelopment, neuroinflammation, and BBB development. The balance between Wnt-β-catenin and Shh pathways controls angiogenesis, barriergenesis, neurodevelopment, central nervous system (CNS) morphogenesis and neuronal guidance. These interactions are critical to maintain BBB function in the mature CNS to prevent influx of pathogens and inflammatory cells. Genetic mutations of key components of these pathways have been identified in ASD patients and animal models which correlate with severity of ASD symptoms. Disruption of the Shh/Wnt crosstalk may therefore compromise BBB development and function. In turn, impaired Shh signaling and glial activation may cause neuroinflammation that could disrupt the BBB. Elucidating how ASD-related mutations of Shh/Wnt signaling could cause BBB leaks and neuroinflammation will contribute to our understanding of the role of their interactions in ASD pathophysiology. These observations may provide novel targeted therapeutic strategies to prevent or to alleviate ASD symptoms while preserving normal developmental processes.

Graphical abstract

Shh-Wnt signaling crosstalk regulates neurodevelopment, central nervous system (CNS) morphogenesis, axon guidance and angiogenesis and plays a role in CNS function maintenance. Genetic mutations of Shh-Wnt signaling components identified in Autism Spectrum Disorders (ASD), a neurodevelopmental disease, correlate with symptoms severity. These mutations result in neuroinflammation, oxidative stress, neurovascular disruption, recruitment of immune cells and platelets, endothelial injury, and the release of inflammatory mediators and serotonin. Identification of mechanisms underlying Shh-Wnt pathway dysregulation in ASD could advance our understanding of its role and provide targeted treatment strategies to prevent or alleviate ASD symptoms while preserving normal development.

1. Introduction

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder described by social communication deficits and limited or repetitive behaviors. Both genetic and environmental factors have been shown to contribute to the development of ASD (Chaste & Leboyer 2012; Sledziowska et al. 2020). Immune dysregulation, oxidative stress and neuroinflammation with glial activation have been proposed to play a role in the pathogenesis of ASD (Vargas et al. 2005; Onore et al. 2012; Anderson & Maes 2014; Bjorklund et al. 2016; Ohja et al. 2018). Neuroinflammation is characterized by activation and proliferation of inflammatory cells, microglia and astrocytes, increased inflammatory cytokines production and neurotrophins dysfunction which can alter immune and synaptic function, brain development and cognitive function (Bjorklund et al. 2016; Ohja et al. 2018; Gomez-Fernandez et al. 2018; Matta et al. 2019). A study of post-mortem human ASD brains showed altered expression of genes associated with blood brain barrier (BBB) integrity with increased neuroinflammation (Fiorentino et al. 2016). BBB breach was also reported in a rat prenatal valproic acid model of ASD (Kumar & Sharma 2016; Kumar et al. 2015). While there are studies reports on gut-brain barrier dysfunction in ASD (Downs et al. 2014; Heberling et al. 2013), more studies need to be conducted to investigate the involvement and the mechanisms of BBB disruption in autism models.

Neuroinflammation and oxidative stress such as those reported in ASD, trigger a cascade of events with dissociation of cell-cell junctions between endothelial cells and cytoskeletal reorganization, resulting in endothelial injury and BBB breach (Kumar et al. 2009; Capaldo & Nusrat 2009; Pun et al. 2009; Daneman & Prat 2015). Moreover, during oxidative stress, activated platelets are recruited to the site of BBB leakage and vascular injury, adhering to the subendothelium, releasing inflammatory mediators-containing granules, and recruiting more inflammatory cells, therefore further escalating the inflammatory process (Figure 1) (Jagadapillai et al. 2016; Mezger et al. 2015; Deppermann 2018).

Figure 1:

Schematic neuroinflammation and vascular interactions in ASD pathogenesis. ASD is characterized by microglial activation causing neuroinflammation with the release of inflammatory mediators and oxidative species. These in turn will further exacerbate microglial activation, and recruit activated leukocytes and platelets, releasing granule content with more inflammatory markers, and serotonin. Endothelial injury and platelet adhesion contribute to blood brain barrier disruption with leakage of endothelial content and fibrinogen, further exacerbating neuroinflammation and glial activation.

The blood brain barrier acts as an active interface between central nervous system and the vascular compartment. The basic unit of the BBB consists of endothelial cell, astrocyte, pericyte, and the surrounding neurons (Hawkins & Davis 2005; Liu et al. 2012), The BBB plays a significant role in regulating cells, ions and molecules exchange of substances, maintaining central nervous system (CNS) homeostasis, controlling the amount of immune cells penetrating the CNS, and protecting the CNS from toxins and pathogens (Daneman & Prat 2015; Sweeney et al. 2019; Kealy et al. 2020). CNS capillaries are lined with specialized endothelial cells held together by tight junctions and containing high amounts of mitochondria critical for active transport processes and for nutrients uptake from the blood (Daneman & Prat 2015; Liu et al. 2012; Langen et al. 2019). The inner and outer basement membranes surround the blood vessel and the end–feet of astrocyte processes ensheathe the vascular tube and the neuronal processes, providing a linkage between neurons and vascular cells (Daneman & Prat 2015; Hawkins & Davis 2005). Failure of the BBB can result from various pathologies such as ischemic stroke, traumatic brain injury or CNS diseases involving neuroinflammation that occurs in multiple sclerosis, encephalitis, Alzheimer Disease, epilepsy (Bell & Zlokovic 2009; Gray & Woulfe 2015; Fiorentino et al. 2016; Kealy et al. 2020; Heinemann et al. 2012; Moor et al. 1994; Hawkins & Davis 2005; Xiao et al. 2020; Milikovsky et al. 2019), and even in aging (Senatorov et al. 2019; Michalicova et al. 2020). BBB dysfunction can also potentially result from genetic abnormalities during development. For instance, recent studies have shown that alterations in the Sonic Hedgehog (Shh) and WNT / β-catenin signaling pathways during development can result in a dysfunctional BBB, leakage of blood toxic components into the CNS, cellular infiltration, defective clearance of molecules, cerebral blood flow reduction and dysregulation, neurological deficits and may be associated with ASD pathogenesis (Liu et al. 2018; Rahi & Mehan 2020; Stenman et al. 2008; Alvarez et al. 2011). Furthermore, although multiple Shh/Wnt and β-catenin pathway gene mutations have been identified as ASD-associated mutations, the role of Shh/Wnt signaling-induced cerebrovascular deficits in ASD pathogenesis has not been elucidated (Caracci et al. 2016; Zhang et al. 2012). Multiple studies have identified BBB dysfunction and neuroinflammation in ASD, however few have examined the implication of Shh/Wnt signaling in combining these associated pathologies. In this review, we will focus on the cross talk between Shh/Wnt signaling and neurovascular signaling which may play an essential role in compromising the development, the integrity, and the function of the BBB, thereby facilitating the development of ASD.

2. Shh/Wnt Signaling and BBB development

BBB formation is a gradual process during embryonic development starting with angiogenesis followed by barriergenesis. The process begins with vascularization of the brain from the perineural vascular plexus at approximately E9 and initiated by the release of growth factors from the neural tube. Blood vessels entering the CNS and coming in contact with astrocytes and pericytes, immediately acquire functional tight junction (TJ) proteins, forming a network, and starting to express various transporters (Obermeier et al. 2013; Liebner & Plate 2010). Activation of the Wnt / β-catenin signaling pathway during angiogenesis induces expression of genes essential for angiogenesis and barriergenesis. Shh, released by astrocytes, signals to endothelial cells to induce TJ proteins and junctional adhesion molecules via Wnt / β-catenin signaling. The Shh signaling pathway also suppresses the expression of proinflammatory factors and intercellular adhesion molecule-1, inhibiting infiltration of immune cells into the CNS (Alvarez et al. 2011; Langen et al. 2019; Sweeney et al. 2019).

The canonical Wnt / β-catenin signaling pathway plays a critical role in embryonic development and drives the angiogenesis process in the CNS but not in non-neural tissue, inducing the CNS-specific vascular system and specialized BBB (Langen et al. 2019; Obermeier et al. 2013; Stenman et al. 2008; Daneman et al. 2009). At least two ligand-receptor signaling complexes, activated in specific brain areas, have been identified, controlling β-catenin stabilization, and activating canonical Wnt signaling, both critical for vascular development. The first pathway, Wnt7a/b ligands signaling through Frizzled (Fzd) receptors and the second, activated by Norrin (encoded by Ndp), a Transcription Growth Factor-β family-related ligand that binds to Fzd4, both acting via β-catenin-dependent transcription, and activating T-cell factor (TCF) transcription (Langen et al. 2019; Daneman et al. 2009; Zhou et al. 2014). The Wnt7a/b pathway is exclusively active in the forebrain and ventral spinal cord, and the Norrin pathway is exclusive to the cerebellum, retina, and dorsal spinal cord while both pathways are active in the brainstem (Cho et al. 2017; Wang et al. 2018; Zhou et al. 2014). The canonical Wnt signaling is the best-characterized BBB formation pathway and is also required to maintain BBB properties in the mature CNS (Langen et al. 2019; Zhou et al. 2014). The expression of certain transporters and TJ proteins such as Glucose transporter-1 and Claudin-5 also correlates with canonical Wnt activation (Langen et al. 2019). Wnt activation is regulated by a family of proteins, the secreted Frizzled-related proteins (SFRPs), that inhibits the Wnt pathway by either directly binding to Wnt or by competing with Wnt for binding to Fzd receptors, disturbing the Wnt-Fzd interaction (Bovolenta et al. 2008). Additionally, Dickkopf (Dkk) proteins, a family of proteins (Dkk 1 – 4), have been identified as Wnt pathway regulators, mostly inhibiting Wnt signaling, however Dkk2 can potentially enhance or inhibit Wnt signaling (Choi et al. 2012; Baetta & Banfi 2019; Chae & Bothwell 2019). Dkk1, the most extensively studied Dkk isoform, binds competitively to Wnt receptors LRP5/6, preventing Wnt binding. Dkk1 expression is regulated by Wnt-dependent TCF transcription, therefore providing a feedback control of its transcription (Choi et al. 2012; Chae & Bothwell 2019). While most studies have examined Dkks in the context of cancer and neurodegenerative disease, recent findings have implicated these proteins in vascular pathologies such as atherosclerosis (Ueland et al. 2009; Baetta & Banfi 2019; Chae & Bothwell 2019). Dkk1 has been shown to play a role in endothelial cells differentiation and angiogenesis most probably through its regulation of Wnt signaling (Choi et al. 2012). In contrast, Dkk2 displays Wnt-independent pro-angiogenic activity, opposite to Dkk1 (Choi et al. 2012). Platelets have been shown as an important source of Dkk1, modulating platelet-endothelial activation, and platelet activation is required for Dkk1 release into the circulation (Ueland et al. 2009; Chae & Bothwell 2019). Studies suggest that platelet originating Dkk1 may play an important immunoregulatory function in inflammatory disease and particularly in vascular inflammation (Chae & Bothwell 2019). Therefore, given the role of the Wnt / β-catenin pathway in BBB development and integrity, it is possible that Dkk proteins may play a role in BBB-related physiopathology and the development of endothelial injury. The Wnt / β-catenin pathway has been shown to interact with multiple pathways such as NOTCH, mammalian target of rapamycin (mTOR) and the Shh pathways (Alvarez-Medina et al. 2008; Ding & Wang 2017; Collu et al. 2014; Borday et al. 2012). One of the SFRP proteins, SFRP1, regulated by Shh-controlled GLI transcription factor, has been identified as a point of convergence in the regulation of Wnt and Shh pathways (Ding & Wang 2017). A schematic representation of these interactions is depicted in Figure 2.

Figure 2:

Shh/Wnt crosstalk signaling in neurodevelopment and blood brain barrier genesis and maintenance. * denotes the gene mutations that have been reported in ASD. APC: Adenomatous Polyposis Coli; DKK: Dickkopf; Fzd: Frizzled; GSK 3β: Glycogen synthase kinas 3β; LRP: LDL receptor-related protein; Shh: Ptch 1: Patched 1; ROZ: Reactive oxygen species; SFRP: Secreted frizzled related protein; Sonic Hedgehog; Smo: smoothened; TCF: T-cell factor; TJ: tight junction; Wnt: Wingless-related integration site.

The Shh pathway, represents a major signaling pathway in the control of embryonic CNS morphogenesis, neuronal guidance, generating distinct classes of neurons within specific dorsal-ventral axis of the developing spinal cord, and has been identified as an important player for the BBB formation (Figure 3) (Alvarez et al. 2011; Obermeier et al. 2013; Nagase et al. 2005; Alvarez-Medina et al. 2008; Liu et al. 2018). Shh deletion in mice is associated with severe BBB abnormalities, decreased expression of tight junction proteins, and embryonic lethality between E11 and E13.5 (Alvarez et al. 2011). When activated, Shh binds and inactivates the receptor Patched-1 (Ptch-1), allowing activation of Smoothened (Smo), activating the Gli family of transcription factors to induce its target genes (Figure 2) (Alvarez et al. 2011; Ding & Wang 2017; Osterlund & Kogerman 2006). Alvarez et al. detected Shh expression in vitro and in situ in human and murine astrocytes but not in endothelial cells or pericytes. In contrast, Ptch-1 and Smo were expressed on BBB endothelial cells but not on astrocytes or pericytes, suggesting that astrocytes use hedgehog signaling to communicate with BBB endothelial cells (Alvarez et al. 2011). Co-culture of non-neural endothelial cells with astrocytes or in astrocyte-conditioned media was shown to induce trans-differentiation of non-neural EC into the brain type, endowing them with BBB properties with increased transporters expression and TJs formation (Hayashi et al. 1997). Deletion of Shh significantly decreased TJ proteins expression, and targeted endothelial Smo deletion resulted in compromised BBB and increased leakage of fibrinogen, immunoglobulins and apolipoproteins. These data all suggest that Shh signaling plays an important role in maintaining BBB properties (Alvarez et al. 2011). Inflammation, induced by treating astrocytes or endothelial cells with tumor necrosis alpha (TNF-α) and interferon-gamma (IFN-γ), increased Shh expression in astrocytes and Ptch-1 and Smo expression in BBB endothelial cells. These observations suggest that the Shh pathway plays a role in regulating BBB function during inflammation and may be acting as an endogenous anti-inflammatory system (Figure 3) (Alvarez et al. 2011). While essential to embryonic vascular development, Shh signaling can be reactivated during vessel repair in adults (Liu et al. 2018; Chechneva et al. 2014). Wnt / β-catenin pathway downregulates Shh by controlling the expression of the main repressor of Shh / Gli signaling, Gli3, however Shh controls Wnt signaling by regulating SFRP1 expression and restricting Wnt target genes expression (Figure 2) (Alvarez-Medina et al. 2008; Ding & Wang 2017). The Wnt antagonist, SFRP1, has been shown to be regulated by Shh (Ding & Wang 2017; Katoh & Katoh 2006) and pharmacological activation of the canonical Wnt pathway inhibited Shh, while a Smo agonist decreased Wnt activity, activating Shh (Ding & Wang 2017; Borday et al. 2012). Overall, these observations indicate that the balance between these pathways regulates neurodifferentiation, proliferation and may be critical for proper BBB development and maintaining its integrity during development and neuroinflammation (Figures 2 and 3). Therefore, given the involvement of the Shh /Wnt pathways in both neurodevelopment and BBB development and maintenance, we hypothesize that therapeutic strategies targeting these pathways may not only prevent vascular abnormalities but also neurodevelopmental pathologies as these appear to share common pathways.

Figure 3:

Physiological and developmental effects of Shh/Wnt cross-regulation. Proper crosstalk of Wnt and Shh pathways ensures well-timed neurodevelopment and function (bottom left part). In contrast, imperfect cross-regulation will induce neuroinflammation, endothelial dysfunction, BBB breach and disrupt neuronal guidance and excitatory/inhibitory balance (bottom right part), all associated with ASD.

ASD: Autism spectrum disorder; BBB: Blood brain barrier; CNS: Central nervous system; DKK: Dickkopf; E / I: excitatory / Inhibitory; SFRP: Secreted frizzled related protein; Sonic Hedgehog; Wnt: Wingless-related integration site.

3. Wnt / β-catenin and Shh signaling pathways in ASD

The canonical Wnt / β-catenin and the Shh pathways, essential for BBB development, have been implicated in genetic models of ASD and in ASD patients as various mutations of key components of these pathways were uncovered (Kalkman 2012; Patel et al. 2017; Kumar et al. 2019; Oron & Elliott 2017). The Wnt and Shh pathways have been shown to cross-regulate, as the induction of the one inhibits the other (Ding & Wang 2017). In addition to its role in angiogenesis and BBB development, Wnt signaling is implicated in neuronal differentiation, morphology, and neurotransmission (Figure 3). Of the multiple gene mutations identified in the Wnt pathways some induce, and others decrease Wnt signaling. However, both hyper- or hypo-activity of the Wnt pathway is deleterious for dendrite growth and will negatively affect cognitive function. Multiple mutations in CTNNB1 gene, encoding β-catenin, have been identified in ASD patients and conditional CTNNB1 KO mice in parvalbumin neurons displayed characteristic ASD core behavioral symptoms (Krumm et al. 2014; Bae & Hong 2018). In a rat model of valproic acid-induced ASD, administration of sulindac, downregulating the Wnt / β-catenin pathway, ameliorated oxidative markers and behavioral deficits (Zhang et al. 2012). These observations suggest that changes in Wnt / β-catenin signaling may also underlie some of the environmentally caused ASD-related symptoms, such as in prenatal valproate exposure. Additional components of the Wnt pathway have been identified as key drivers of ASD (Kalkman 2012; Kumar et al. 2019; Oron & Elliott 2017). Hyperactivation of GSK3, a negative regulator of β-catenin, have been detected in several brain regions of a Fragile X syndrome) mouse model. Impaired social preference and exaggerated anxiety during social interaction have been identified in a GSK3 knock in mouse model, while inhibition of GSK3 in the hippocampus of Fragile X syndrome mice rescued behavioral deficits (Oron & Elliott 2017; Caracci et al. 2016; Min et al. 2009; Franklin et al. 2014). Deletion of the GSK3 inhibitor “disrupted in schizophrenia 1”, DISC1, was found in BTBR mice, an inbred model of ASD. In addition, recent studies in mouse ASD models have determined how autism-related mutations may induce autism-related behaviors through the dysregulation of Wnt-signaling. One such example is the chromodomain-helicase-DNA-binding protein 8 (CHD8) gene, that was recently found to be strongly associated to ASD, and encodes for a chromodomain helicase DNA binding protein that acts as a positive regulator of the Wnt pathway in the brain and binds to the promoter regions of Fzd1, Dishevelled-3 and β-catenin (Oron & Elliott 2017; Durak et al. 2016). CHD8 knockdown during cortical development results in defective brain development. Expressing a degradation-resistant β-catenin construct in CHD8-downregulated embryos, rescued aberrant dendritic arborization as well as increased spine density and restored the aberrant social and anxiogenic behaviors to the levels seen in control mice (Oron & Elliott 2017). Mutations in FZD9 gene, a receptor for the canonical ligand Wnt2, and deletion or multiplication of its downstream signaling partner B-Cell Lymphoma 9 Protein and of a β-catenin co-factor, CREB Binding Protein, have been reported in ASD patients (Kalkman 2012). Additionally, TCF mutations have been found in ASD patients. TCF20 KO impairs neurogenesis and induces autistic-like (Feng et al. 2020)behavior in mice, that can be rescued by overexpressing TCF4 (Feng et al. 2020).These studies suggest a direct link between Wnt-signaling and ASD-like behavior in autism mouse models.

Shh signaling is critical in the developing as well as in the adult brain function (Patel et al. 2017). Correlation of Shh pathway and its signaling proteins with ASD etiology has been well established, with increased Shh proteins serum levels in ASD children correlating with symptoms severity (Baranova et al. 2020). Shh signaling plays an antioxidant and survival role in ASD, increasing superoxide dismutase and glutathione peroxidase and suppressing pro-apoptotic signaling (Patel et al. 2017; Ghanizadeh 2012). Shh activation causes its binding to Ptch receptors and activates Gli proteins translocation to the nucleus, activating transcription targets and repressing the Wnt / β-catenin signaling (Figure 2) (Alvarez-Medina et al. 2008; Ding & Wang 2017; Katoh & Katoh 2006; Kumar et al. 2019). PTCH1 gene has been shown to be upregulated in ASD children, but downregulated in adults (Benitez-Burraco et al. 2016). Ptch1 overexpression and over activity may depress the Shh pathway, contributing to the severity of ASD symptoms (Baranova et al. 2020). Mutations in PTCHD1 (Patch domain containing 1), a structurally ptch1-related gene, have been reported in some ASD patients and their family and absence of ptchd1 in mice shows neurodevelopmental disorders and excitatory / inhibitory imbalance (Chaudhry et al. 2015; Torrico et al. 2015; Tora et al. 2017; Noor et al. 2010; Ung et al. 2018). Additionally, dysregulation of Shh in the brain has been implicated in ASD cell fate, axon guidance, patterning and regionalization (Patel et al. 2017; DeRosa et al. 2018; Kumar et al. 2019). In a study of axon guidance in spinal cord midline crossing, Parra and Zou reported that blocking Shh could cause severe guidance defects at different stages of development, resulting in phenotypes reminiscent of that of the semaphorin 3F (Sema 3F) receptor neuropilin2 (NRP2) deficient mice (Parra & Zou 2010). Shh can activate the pre-crossing repulsive response to Sema 3B and 3F, in spinal cord commissural axons midline crossing and regulate growth cone sensitivity to semaphorins (Parra & Zou 2010). Interestingly, we have previously characterized NRP2 knockout mice as a novel murine genetic model of autism and epilepsy (Gant et al. 2009) and reported that selective deletion of the NRP2 ligand, Sema 3F, in GABAergic interneurons is associated with decreased interneurons numbers, oxidative stress, and also exhibited ASD behavioral symptoms (Li et al. 2019).

Shh and Wnt morphogens both play attractive or repulsive axon guidance roles in different contexts of neurodevelopment, and therefore could also underlie the mechanisms of Sema3F/NRP2 KO-mediated development of ASD symptoms. A time course transcriptome analysis of genes differentially expressed between ASD and control iPSC-derived neurons, attempting to replicate the neurodevelopmental process, shows evidence for disturbance in pathways linked to synaptic function, axon path-finding, cell fate specification, and activity-dependent regulation of gene expression, such as the Shh / Gli1 signaling pathway (DeRosa et al. 2018). Identifying the temporal sequence of these signaling pathways is essential to evaluate their contribution to the pathogenesis and physiological evolution of ASD symptoms. Overall, it is well established that ASD is associated with a wide range of genetic mutations. Multiple among these mutations affect the Shh/Wnt pathway, such as mutations in the Wnt ligand family (Wnt 1, 2, 3, 7a, 7b, 9b, etc.…) and of well documented Wnt signaling-related molecules mutations such as CTNNB1(β-catenin), FzD9, APC, GSK3β, DISC, CH8D, TCF. In addition, genetic changes altering Shh, PTCH1, PTCHD, and Shh-dependent guidance molecules such as Sema 3F-NRP2, have been shown to be associated with ASD symptoms. All the above mutations are also substantially implicated in vascular development and in abnormal BBB development and maintenance. Therefore, more studies are required to identify how Shh/Wnt pathways interactions affect neurovascular development and neurogenesis processes. The crosstalk between these morphogenic pathways may underlie the differences between normal development and ASD physiopathology and suggest a prominent role of endothelial abnormalities in the development of the disease.

4. Neurovascular Signaling in ASD

The biological processes are similar in vascular and neuronal development in the brain. Generation of different types of cells begins with the proliferation of stem cells. Common cell cycle mechanisms regulate the proliferation of angioblasts and neuronal precursors. In both systems, excess cell elements are produced, guided to the appropriate placement, and then sculpted by the processes of apoptosis and pruning. Each system is driven to undergo activity-dependent (either shear force of blood flow or electrical impulses of neuronal networks) re-modeling during development. Furthermore, both developing vascular and neural tissue must exhibit remarkable plasticity to respond to changing environments. Correct CNS process outgrowth, guidance, and placement involve neuronal migration, axon/dendritic growth cone guidance, pathfinding, branching/arborization, and synaptogenesis (Baruah et al. 2019). The developing vascular system has filopodial extensions of tip cells, endothelial cell migration, vessel elongation, and sprouting. The common mechanisms regulating the genesis of endothelium and neurons in the CNS are confirmed by the overlapping repertoire of signaling molecules and their regulation by compartment specific transcription factors (Shh/Wnt). These common systems regulate angiogenesis, neurogenesis, and neuronal migration in the brain. Importantly, molecules produced in one cell type influence the other to promote proliferation, differentiation, migration, or process outgrowth in both systems.

As an example, during development, preformed vascular networks serve as the cellular substrate for the long-distance migration of GABAergic neurons (important cell types in ASD and epilepsy) (Datta et al. 2020). In particular, periventricular endothelial cells of the embryonic forebrain have a unique gene expression signature compared to pial endothelial cells or endothelial cells from midbrain or hindbrain (Choi & Vasudevan 2019). Novel work has shown that GABA_A receptor-GABA signaling pathway in perivascular endothelial cells of the embryonic forebrain, has a distinct impact on GABAergic cell biology. Knockout of perivascular endothelial cell GABA signaling results in mice with both seizures and autistic behaviors (Li et al. 2018). The absence of endothelial GABA release, affects to some degree all key cellular events during forebrain development, including angiogenesis, neurogenesis, radial migration of projection neurons, and tangential migration of GABAergic neurons, (Li et al. 2018). Similarly, another neurovascular signaling system, the semaphorin-neuropilin system, alters excitatory and GABAergic cell development, vascular development, and this dysfunction results in seizures and autistic behaviors, reminiscent of the changes after Shh inhibition (Gant et al. 2009; Li et al. 2019; Geretti et al. 2008; Parra & Zou 2010). Multiple semaphorin ligands, including semaphorin 3F (Sema 3F), Sema 3A, and Sema 4D, and Sema 7A, may impact blood brain barrier properties and regulate the flow of immune cells across the blood brain barrier (Zhang et al. 2020; Le Guelte et al. 2012; Smith et al. 2015). These findings strongly support the involvement of common signaling networks such as Shh/Wnt, interacting to regulate vascular and neuronal ontogenesis, and therefore could explain the interdependence of these developmental processes. As depicted in Figure 3, we hypothesize that dysfunctional interactions between Shh/Wnt signaling and neurovascular signaling may impair BBB function and induce neuroinflammation as a pathophysiological process in ASD.

5. BBB, Neuroinflammation and ASD

As discussed above, mutations of Shh/Wnt signaling correlate with the severity of core ASD symptoms. As summarized in Figure 3, Shh has also been shown to trigger antioxidant pathways and regulate inflammation. Disrupting Shh/Wnt crosstalk may therefore impair neurodifferentiation and CNS morphogenesis, and perturb the formation of a functional BBB. At later developmental stages disruption of Shh/Wnt pathways has been shown to affect the maintenance of BBB integrity, the regulation of inflammatory pathways and endothelial function, leading to persistent neuroinflammation and impaired excitatory/inhibitory balance that have been shown to occur in ASD. Thus, Shh/Wnt signaling may perturb BBB function inducing neuroinflammation. Alternatively, impaired Shh signaling could trigger oxidative stress and neuroinflammation resulting in BBB disruption, and intensify neuroinflammation. Determining how neuroinflammation and BBB disruption interact to generate ASD-related symptoms would provide novel strategies to alleviate these symptoms.

Neuroinflammation is an essential component of the CNS innate immunity to restrain infection and eliminate pathogens, cell debris, and misfolded proteins. Neuroinflammation initially plays a part in neural tissue repair and injury resolution, however in chronic neurological diseases, it becomes persistent and damaging to the CNS (Yang & Zhou 2019; Shabab et al. 2017). Chronic activation of glial cells increases the release of pro inflammatory cytokines, such as TNF-α, IL1β, IL6, IFN-γ and reactive oxygen and nitrogen species that induce neurodegeneration and neuronal dysfunction (Yang & Zhou 2019; Shabab et al. 2017). The joint activation of multiple glial cell types and peripheral immune cells such as neutrophils, T cells, and macrophages, results in recruitment of additional blood borne immune cells can be activated and interact with CNS blood vessels, increasing vascular permeability, releasing reactive oxygen species, and disrupting the BBB (Daneman & Prat 2015). Various neurodegenerative diseases, whether sporadic or linked to genetic mutations are associated with neuroinflammation and BBB breakdown (Kealy et al. 2020; Xiao et al. 2020; Noe et al. 2020; Sweeney et al. 2019). While occurring in multiple neurological diseases, neuroinflammation effect on BBB integrity have been extensively studied in Alzheimer Disease (Michalicova et al. 2020; Milikovsky et al. 2019; Noe et al. 2020; Montagne et al. 2017). A critical player in BBB disruption is uncontrolled microglial activation.

Microglial cells are resident CNS parenchymal immune cells that are involved in regulating neuronal development, innate immune response, and wound healing and can act as antigen-presenting cells in adaptive immunity. In addition to their role in CNS maintenance, microglia sense and rapidly respond to changes in CNS environment. Microglia are reactive to their environment, change their morphology to acclimatize to their microenvironment, to neuronal activity, and to synaptic signaling (Domercq et al. 2013; Fernandez-Arjona et al. 2017; Miyamoto et al. 2013; Olah et al. 2011). Throughout neuroinflammation, microglia endure de-branching and can either become cells with larger cell body and thicker processes or can get small spherical or rod like shapes (Davis et al. 1994). Different studies have attempted to link neuroinflammation and the morphological changes in microglia with pathophysiological characteristics of brain disease (Fernandez-Arjona et al. 2017; Baron et al. 2014).

Activated microglia have been identified in ASD and release a variety of inflammatory mediators, cytokines and chemokines such as IL-1, IL-2, IL-6, Tumor Necrosis Factor alpha, IFN-γ, and others which in turn trigger astrocyte reactivity (Ohja et al. 2018; West et al. 2019; Vargas et al. 2005; Balasingam et al. 1994). As a protective mechanism both activated microglia and astrocytes form a glial scar at the injury site to stop additional damage and stimulate tissue repair (Adams and Gallo, 2018). Reactive astrocytes can further recruit microglia (Rolls et al., 2008), and with continued release of pro-inflammatory molecules as well as reactive oxygen species, can induce vasculo-endothelial dysfunction and structural impairment to the tissue in the surrounding brain region (Mittal et al. 2014). Vascular Inflammation may trigger prothrombotic events, upregulation of fibrinogen, recruitment of inflammatory cells and platelets, and endothelial cells activation, contributing to persistent vascular inflammation, redox imbalance, pro-leaks molecules and recruiting inflammatory mediators to the endothelium (Figure 1) (Dixon et al. 2012). Fibrinogen, essential to vascular homeostasis and a main component of blood clots, is usually excluded from the brain parenchyma by the blood brain barrier. Presence of fibrinogen, suggestive of BBB disruption and platelet aggregation, was reported in the brain of neurodegenerative patients mainly accumulating in neurovasculature and attracting more platelets and microglia (Ahn et al. 2010; Cortes-Canteli et al. 2010; Ryu & McLarnon 2009). Therefore, in pathological conditions, a ruptured BBB could let plasma proteins such as fibrinogen leak into the brain parenchyma, attracting more inflammatory cells and platelets releasing inflammatory granules content and activating endothelial cells, contributing to persistent vascular inflammation and leakage (Figure 1) (Kumar et al. 2009; Daneman & Prat 2015; Adams et al. 2007; Ed Rainger et al. 2015). Platelet-endothelial activation has been described in several neurological diseases including in ASD patients (Ciabattoni et al. 2007; Bijl et al. 2015; Sevush et al. 1998; Yao et al. 2006). The Wnt / β–catenin pathway, essential to neurodevelopment and BBB maintenance, has been reported to be functional in platelet and regulate their function (Kumari & Dash 2013; Steele et al. 2009). The platelet-derived Wnt inhibitor DKK1 has been shown to contribute to platelet-endothelial interaction (Baetta & Banfi 2019). Additionally, circulating Shh has been reported to inhibit platelet apoptosis, increasing their survival, and to regulate their activity in various diseases conditions (Kumari & Dash 2013). Therefore, Shh-mediated inhibition of platelet removal and increased platelet release of DKK1 could contribute to impair endothelial function and promote BBB breach. These elements may propagate persistent cerebrovascular inflammation resulting in endothelial injury and breach of the BBB.

ASD pathogenesis is associated with neuroinflammation (Anderson & Maes 2014; Matta et al. 2019; Nakagawa & Chiba 2016; El-Ansary & Al-Ayadhi 2012; Kern et al. 2015). Human brain samples from ASD patients show evidence of microglial activation, ongoing neuroinflammatory processes in different brain regions, and increased reactive microglia presence also noted in the dorsolateral prefrontal cortex (Vargas et al. 2005; Pardo et al. 2005; Takano 2015; Zimmerman et al. 2005; Morgan et al. 2010; Tetreault et al. 2012). Studies reported widespread microglial activity in cortical regions, white matter and the cerebellum of postmortem ASD patients when compared to controls, and elevated astrocyte glial fibrillary acidic protein immunoreactivity in different brain areas (Vargas et al. 2005; Morgan et al. 2010; Laurence & Fatemi 2005; Edmonson et al. 2014; Ahlsen et al. 1993; Rosengren et al. 1992). Astrocytes play a central role in the development of synapses and regulation of neuronal processes that control circuit formation (Clarke & Barres 2013; Mederos & Perea 2019). Additionally, astrocytes express GABA and glutamate receptors and transporters, regulating glutamate release and re-uptake, neuronal energy and metabolism, and synapse function, all dysregulated in ASD (Mederos & Perea 2019; Belanger et al. 2011). Therefore, astrocyte activation may contribute to the excitatory/inhibitory imbalance characteristic of ASD (Kuo & Liu 2018; Nelson & Valakh 2015).

BTBR mice, an inbred mouse model of ASD, exhibit a neuroinflammatory profile that correlate with some of their repetitive behavior, characteristic of ASD (Careaga et al. 2015; Onore et al. 2013). We have previously described a neuropilin-2 knockout as a novel murine genetic model of autism and epilepsy (Gant et al. 2009). Furthermore, we reported that selective deletion of the NRP2 ligand, Sema 3F, in GABAergic interneurons is associated with decreased interneurons numbers, exhibited ASD behavioral symptoms, and showed microglial activation and increased inflammatory and oxidative markers in several brain regions (Li et al. 2019). Sema 3F is a neuronal guidance molecule in the developing brain and NRP2 / Sema 3F signaling regulate axon patterning, GABA / excitatory synaptogenesis, cell and neurite motility of the neurons and endothelial cells, and is critical to neurogenesis and vasculogenesis (Sahay et al. 2003; Alto & Terman 2017; Raimondi & Ruhrberg 2013; Gilbert & Man 2017). Sema 3F KO mice exhibit decreased locomotor activity and abnormal anxiety and fear-related behavior (Matsuda et al. 2016). In the adult brain, Sema 3F / NRP2 signaling plays a role in the regulation of hippocampal synaptic transmission and NRP2-deficient mice show impaired hippocampal-dependent behavior (Sahay et al. 2003; Sahay et al. 2005; Shiflett et al. 2015). Endothelium-derived Sema 3G signaling via NRP2 regulates hippocampal structure and plasticity (Tan et al. 2019). These pathways also play a role in immune modulation and in the resolution of inflammation (Mirakaj & Rosenberger 2017). Neuropilins and class 3 semaphorins have been shown to be critical to embryonic angiogenesis (Staton 2011; Takashima et al. 2002). In a model of acute inflammation, NRP-2 deficiency increased vascular permeability and decreased lymphatic capillaries, resulting in severe swelling and lymphedema and addition of Sema 3F prevented increased vascular permeability and loss of lymphatic drainage, critical to limiting inflammation (Huggenberger et al. 2011; Mucka et al. 2016). Sema 3F has also been shown to prolong platelets lifespan, and increased platelet counts have been reported in ASD (Bijl et al. 2015; Kumari & Dash 2013). Increased platelet numbers in ASD patients may enhance platelet aggregation and endothelial injury, compromising BBB integrity. Platelets alpha granules comprise several important proteins such as fibrinogen, P-selectin, von Willebrand factor, growth factors, while dense granules release glutamate and serotonin, one of the first ASD biomarkers identified (Blair & Flaumenhaft 2009; Sharda & Flaumenhaft 2018; Padmakumar et al. 2019). Hyperserotonemia in blood and in platelets was detected in 17 % to 40 % of ASD patients and correlate with increased behavioral impairment (Padmakumar et al. 2019; Muller et al. 2016). Furthermore, our preliminary studies using the Sema 3F-KO mouse model of ASD show vascular injury, with increased fibrinogen, serotonin, and albumin leakage, suggestive of platelet activation and of disrupted BBB. Thus, NRP2 and Sema 3F deletion in mice may affect not only CNS development, neurite growth and axonal guidance but also dysregulate immune and inflammatory responses, and vascular development, all potentially contributing to BBB dysfunction and reported in ASD.

5. Conclusions

The crosstalk of Wnt / Shh pathways and vascular systems control neurovascular unit development, regulate neural progenitor cells differentiation, neuronal migration, axon guidance, dendritogenesis, synaptogenesis, BBB formation and function. These interactions, depicted in Figure 3, are important for neuronal plasticity during development and in adult organisms. Disruptions in their proper functioning and interactions can contribute to neurodevelopmental disorders, including ASD. Numerous therapeutic interventions, targeting different element of their signaling pathways and their regulation are already in trial or in use for the treatment of several neurological diseases (Rahi & Mehan 2020; Baranova et al. 2020). While modulation of the Wnt-β-catenin and Shh pathways may offer potential therapeutic benefit in the management or the prevention of ASD symptoms, they require better understanding of their crosstalk during neurovascular and neurological development, as well as time and spatial fine-tuning. Indeed, changes in Wnt and Shh expression and activity have been shown to affect differently various cell types and brain areas and the timing of these interventions is critical for preserving normal neurovascular development.

One possibility during brain development is that Shh/Wnt signaling regulates the expression of neurovascular signaling like the Sema 3F-NRP2 system and thus may be a set of downstream therapeutic targets. Shh signaling is required for expression of NRP2 mediated cardiac morphogenesis and induces semaphorin 3F repulsion of commissural axons during midline crossing of fibers in the spinal cord (Parra & Zou 2010). Semaphorin 3F is known to regulate GSK3 and β-catenin through NRP2 signaling (Liu et al. 2016; Rao et al. 2015; Samuel et al. 2011; Ng et al. 2016). Shh mediates protective pathways in stroke, regulating microglia/macrophages and inflammatory cytokines and has been implicated in the BBB response to neuroinflammation (Alvarez et al. 2011). Therefore, there is a high likelihood of crosstalk between Shh/Wnt and Sema-NRP signaling pathways. Targeting specific regulators of these pathways rather than the whole pathway, either during development or at later stages, will also allow more selective outcomes, avoiding disruption of the normal development and aiming at ASD treatment and maintenance of the CNS. These and other novel biologicals, targeting the regulation of Wnt -Shh pathways or downstream targets, could alleviate some of ASD symptoms.

List of abbreviations:

ASD

Autism Spectrum Disorder

BBB

Blood Brain Barrier

CNS

Central Nervous System

CHD8

Chromodomain-helicase-DNA-binding protein 8

DKK

Dickkopf

Fzd

Frizzled

GABA

Gamma aminobutyric acid

GSK3

Glycogen synthase kinase 3

IFNγ

Interferon gamma

KO

knockout

NRP2

Neuropilin 2

Sema 3F

semaphorin 3F

Ptch-1

Patched −1

SFRP

Secreted Frizzled Related Protein

Shh

Sonic Hedgehog

Smo

Smoothened

TJ

tight Junction

TCF

T-Cell Factor

TNF-α

tumor necrosis alpha

Wnt

Wingless-related integration site

APPENDIX

--Human subjects --

Involves human subjects:

If yes: Informed consent & ethics approval achieved:

=> if yes, please ensure that the info “Informed consent was achieved for all subjects, and the experiments were approved by the local ethics committee.” is included in the Methods.

ARRIVE guidelines have been followed:

No

=> if it is a Review or Editorial, skip complete sentence => if No, include a statement in the “Conflict of interest disclosure” section: “ARRIVE guidelines were not followed for the following reason:

Not applicable - this is a review no experiments were performed “

(edit phrasing to form a complete sentence as necessary).

=> if Yes, insert in the “Conflict of interest disclosure” section:

“All experiments were conducted in compliance with the ARRIVE guidelines.” unless it is a Review or Editorial